J Physiol 582.1 (2007) pp 73–85 73

Vitamin B1 (thiamine) uptake by human retinal pigmentepithelial (ARPE-19) cells: mechanism and regulation

Veedamali S. Subramanian1, Zainab M. Mohammed1, Andres Molina1, Jonathan S. Marchant2,

Nosratola D. Vaziri1 and Hamid M. Said1,3

1Departments of Medicine (Nephrology) and Physiology/Biophysics, University of California, Irvine, CA 92697, USA2Department of Pharmacology, University of Minnesota Medical School, Minneapolis, MN 55455, USA3Department of Veterans Affairs Medical Center, Long Beach, CA 90822, USA

Retinal abnormality and visual disturbances occur in thiamine-responsive megaloblastic

anaemia (TRMA), an autosomal recessive disorder caused by mutations in the human thiamine

transporter-1 (hTHTR-1). Human retinal pigment epithelial cells play a pivotal role in supplying

thiamine to the highly metabolically active retina but nothing is known about the mechanism,

regulation or biological processes involved in thiamine transport in these cells. To address these

issues, we used human-derived retinal pigment epithelial ARPE-19 cells to characterize the

thiamine uptake process. Thiamine uptake is energy- and temperature-dependent, pH-sensitive,

Na+-independent, saturable at both the nanomolar (apparent K m, 30 ± 5 nM) and the

micromolar (apparent K m, 1.72 ± 0.3 μM) concentration ranges, specific for thiamine and

sensitive to sulfhydryl group inhibition. The diuretic amiloride caused a concentration-

dependent inhibition in thiamine uptake, whereas the anti-trypanosomal drug, melarsoprol,

failed to affect the uptake process. Both hTHTR-1 and hTHTR-2 are expressed in ARPE-19 cells

as well as in native human retinal tissue with expression of the former being significantly higher

than that of the latter. Uptake of thiamine was adaptively regulated by extracellular substrate

level via transcriptionally mediated mechanisms that involve both hTHTR-1 and hTHTR-2; it

was also regulated by an intracellular Ca2+–calmodulin-mediated pathway. Confocal imaging of

living ARPE-19 cells expressing TRMA-associated hTHTR-1 mutants (D93H, S143F and G172D)

showed various expression phenotypes. These results demonstrate for the first time the existence

of a specialized and regulated uptake process for thiamine in a cellular model of human retinal

pigment epithelia that involves hTHTR-1 and hTHTR-2. Further, clinically relevant mutations

in hTHTR-1 lead to impaired cell surface expression or function of the transporter in retinal

epithelial ARPE-19 cells.

(Resubmitted 22 January 2007; accepted after revision 23 April 2007; first published online 26 April 2007)

Corresponding author H. M. Said: Department of Veterans Affairs, Medical Center-151, Long Beach, CA 90822, USA.

Email: hmsaid@uci.edu

The retinal pigment epithelium is part of the blood–retinalbarrier and plays an important role in the delivery ofnutrients to the highly differentiated and metabolicallyactive retina (Pow, 2001; Philp et al. 2003). Cells of thisepithelium form a single layer that is located between thechoriocapilliaris and the retina, and thus, impairment inthe nutrient transport function of these cells may adverselyaffect the retina. Like all other cells of the body, cellsof the retina cannot synthesize the water-soluble vitaminB1 (thiamine), and must obtain the micronutrient fromexogenous sources. The importance of thiamine (in its

V. S. Subramanian and Z. M. Mohammed contributed equally to this

work.

pyrophosphate form) for cellular functions stems from therole it plays as a cofactor for transketolase, α-ketoglutaratedehydrogenease and pyruvate dehydrogenase, enzymesimportant in carbohydrate metabolism. Further, becausethiamine bridges the glycolytic and the pentose phosphatemetabolic pathway that is critical for creating chemicalreducing power in cells, the vitamin is also consideredas having a role in reducing oxidative stress (Calingasanet al. 1996, 1997; Frederikse et al. 1999). Thus, lowintracellular levels of thiamine will lead to impairment inenergy metabolism and a propensity for oxidative injury.In addition, a reduction in intracellular thiamine hasbeen shown to lead to apoptotic cell death (Stagg et al.1999). Clinically, thiamine deficiency (which occurs indisease conditions such as alcoholism, diabetes mellitus

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society DOI: 10.1113/jphysiol.2007.128843

74 V. S. Subramanian and others J Physiol 582.1

Table 1. Gene-specific primers used for generating mutations in SLC19A2

Amino acid Forward and reverse primers 5′–3′

hTHTR-1 CCGCTCGAGATGGATGTGCCCGGCCCGGGATCCTGAAGTGGTTACTTGAGAACT

D93H GTTTCCTGTGTTCCTTGCCACACACTACCTCCGTTATAAACGTTTATAACGGAGGTAGTGTGTGGCAAGGAACACAGGAAAC

S143F ATTGCCTATTACTTTTATATCTACAGTGTGGTGGACCTGGGCGCCCAGGTCCACCACACTGTAGATATAAAAGTAATAGGCAAT

G172D GGTGGGCTTTACAGTGGACTCTGTCCTAGGGCAAATCCGGATTTGCCCTAGGACAGAGTCCACTGTAAAGCCCACC

The table shows the required nucleotide changes in forward and reverseprimers (bold face) used for generation of point mutations. Restriction sitesfor XhoI (italics) and BamHI (underlined) were incorporate into the SLC19A2primer to aid subcloning into the GFP–N3 vector.

and coeliac disease; Leevy & Baker, 1968; Tomasuloet al. 1968; Saito et al. 1987; Victor et al. 1989;Tallaksen et al. 1992) leads to a plethora of abnormalitiesthat include cardiovascular and neurological disorders(Victor et al. 1989; Tanphaichirt, 1994; Berdanier, 1998).Impaired cellular thiamine accumulation that occursin the autosomal recessive disorder thiamine-responsivemegaloblastic anaemia (TRMA), leads to (among otherthings) abnormalities in the retina and visual disturbances(Raz et al. 2000; Meire et al. 2000; Scharfe et al.2000). TRMA is caused by mutational defects inthe human thiamine transporter-1 (hTHTR-1) (seebelow). In contrast to the negative effects that thiaminedeficiency causes in occular (and other) tissues, thiaminesupplementation appears to be of potential benefit inpreventing diabetic retinopathy (Hammes et al. 2003).

How cells obtain thiamine from their surroundingenvironment has been the subject of great interestover the past two decades. Two specific thiamineuptake systems have been identified, namely the humanthiamine transporter-1 and -2 (hTHTR-1 and hTHTR-2,the products of the SLC19A2 and SLC19A3 genes,respectively), and are believed to play important roles inthiamine import to and across cells (Diaz et al. 1999; Duttaet al. 1999; Fleming et al. 1999; Labay et al. 1999; Eudyet al. 2000; Rajgopal et al. 2001; Said et al. 2004). Thepattern of expression of these two thiamine transporters,their regulation, and the degree to which they contributeto cellular thiamine uptake, show some level of tissuespecificity (Diaz et al. 1999; Fleming et al. 1999; Labayet al. 1999; Eudy et al. 2000; Reidling et al. 2002; Saidet al. 2004). However, the mechanism by which retinalpigment epithelial cells transport thiamine and whetherthe process is regulated by intracellular and extracellularfactors is not known. Addressing these issues is importantfrom a physiological and nutritional prospective, andinformative regarding the understanding of pathologicaldisorders encountered in TRMA. Therefore, we used thehuman-derived retinal pigment epithelial (ARPE-19) cell

line, as well as native human retina in our investigationstowards this goal. Our choice of this cell line was based onthe proven similarities between this and cell lines of thenative human retinal pigment epithelial (hRPE), and itsdemonstrated suitability for physiological investigationsthat resulted in similar findings to those occurring in vivo(Dunn et al. 1996; Aukunuru et al. 2001; Busik et al. 2002).Results of our investigations showed for the first time theexistence of a specialized uptake process for thiamine that ispH-dependent, Na+-independent and involves the activityof both hTHTR-1 and hTHTR-2. The results also showedthe process to be adaptively regulated by the prevailingsubstrate level (via what appears to be a transcriptionallymediated mechanism that involves both hTHTR-1 andhTHTR-2), and by an intracellular Ca2+–calmodulin(CaM)-mediated pathway. Finally, our investigations alsoresolved the cellular targeting in ARPE-19 cells of threehTHTR-1 mutations found in TRMA patients.

Methods

Materials

Green fluorescent protein vector (GFP-N3) and humanretinal RNA were obtained from Clontech (Palo Alto,CA, USA). ARPE-19 cell line was obtained from ATCC(Manassas, VA, USA). DNA oligonucleotide primers(Table 1) were from Sigma Genosys (Woodlands, TX,USA). [3H]-Thiamine (specific activity, 10 Ci mmol−1;radiochemical purity, > 98%) was obtained fromAmerican Radiolabeled Company (ARC) (St Louis, MO,USA). Unlabelled thiamine, and all other biochemicalsand molecular biology reagents were purchased fromcommercial sources and were of analytical grade.Melarsoprol (Mel B) and melarsenoxide (Mel Ox) werekindly provided by Dr Reta Brun of the Swiss TropicalInstitute, Basel, Switzerland. Mel B (90 mmol l−1) wasdissolved in propylene glycol and Mel Ox (25 mmol l−1)in dimethyl sulfoxide (DMSO) as stock solutions.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 75

Cell culture, uptake assays and transient transfection

ARPE-19 cells were maintained in Dulbecco’s modifiedEagle’s medium (DMEM) containing 1 μm thiamine.Medium was supplemented with 10% fetal bovineserum (FBS), glutamine (0.29 g l−1), sodium bicarbonate(3.7 g l−1), penicillin (100 000 units l−1) and streptomycin(10 mg l−1). Routine [3H]-thiamine uptake assaywas performed on confluent monolayers (3–4 daysafter confluence) of ARPE-19 cells grown on solidsupport according to a previously established procedure(Aukunuru et al. 2001; Busik et al. 2002; Said et al.2005). To determine the functionality of hTHTR-1clinical mutants, ARPE-19 cells were transientlytransfected with hTHTR-1–GFP, hTHTR-1[D93H]–GFP,hTHTR-1[S143F]–GFP, hTHTR-1[G172D]–GFP andGFP vector alone. Uptake assay was performed after24 h. Protein concentrations were estimated in parallelwells using a protein assay kit (Bio-Rad, CA, USA). Forimaging transient transfection, cells were grown on sterileglass-bottomed Petri dishes (MatTek, MA, USA) andtransfected at 90% confluency with 2 μg plasmid DNAusing Lipofectamine 2000 (Invitrogen, CA, USA). After24 h, cells were analysed by confocal microscopy.

In the studies to investigate the effect of growingthe ARPE-19 cells in the presence of differentexogenous thiamine levels on [3H]-thiamine uptakeand related parameters, the cells were grown for5 days in custom-made thiamine-deficient DMEM (LifeTechnologies Inc., MD, USA) supplemented with 2.5%dialysed FBS (Hyclone). Various levels of thiamine werethen added to the culture medium as specified. Todetermine the degree of thiamine metabolism after uptakeby ARPE-19 cells, a thin-layer chromatography (TLC)procedure employing cellulose gel-precoated plates anda solvent system of isopropanol/0.5 m acetate buffer(pH 4.5)/water (65/15/20, v/v/v) was used as previouslydescribed (Said et al. 2001).

Semiquantitative and real-time PCR analysis

Total RNA (5 μg) was isolated from ARPE-19 cellsand primed with oligo-dT primers to synthesize firststrand cDNA (Superscript; first strand synthesis RT-PCR kit, Invitrogen). To amplify the coding regionof hTHTR-1, hTHTR-2 and β-actin, we used thefollowing gene specific primers for semiquantitative PCRprimers: hTHTR-1: forward, 5′-GCGTCGACGCCAGT-GGCCTTGGATTA-3′; reverse, 5′-CGGGATCCTGAA-GTGGTTACTTGAGAACT-3′; hTHTR-2: forward,5′-GCGTCGACCAGAGAGGGCTCAACT-3′; reverse, 5′-CGGGATCCGAGTTTTGTTGACATGATGATATTAC-3′;β-actin: forward, 5′-TTGTAACCAACTGGGACGATAT-GG-3′; reverse, 5′-GATCTTGATCTTCATGGTGCTAGG-3′, and real-time PCR primers: hTHTR-1: forward,

5′-AGCCAGACCGTCTCCTTGTA-3′; reverse, 5′-TAGA-GAGGGCCCACCACAC-3′; hTHTR-2: forward, 5′-TTC-CTGGATTTACCCCACTG-3′; reverse, 5′-GTATGTCC-AAACGGGGAAGA-3′; β-actin: forward, 5′-CATCCT-GCGTCTGGACCT-3′, reverse, 5′-TAATGTCACG-CACGATTTCC-3′. Real-time PCR conditions were aspreviously described (Said et al. 2004; Reidling et al.2006) and data were normalized to β-actin, and thenquantified using a relative relationship method suppliedby the iCycles manufacture (Bio-Rad). The sample withthe lowest expression level is assigned an arbitrary value ofone and the levels of the rest of the samples are expressedrelative to the expression level above that sample. Eachunit a given gene is expressed above the basal sampleindicates a doubling of the expression level (Livak &Schmittgen, 2001; Reidling et al. 2006).

Analysis of hTHTR-1 and hTHTR-2 expression inARPE-19 cells by Western blotting

Membranous fractions from ARPE-19 cells were iso-lated by mechanical homogenization of cells in a buffercontaining (mm): mannitol 300, EGTA 5 and Tris-HCl12, and a cocktail of protease inhibitors (BoehringerManheim, NJ, USA). Western blot analysis was performedas we previously described using specific anti-hTHTR-1and anti-hTHTR-2 polyclonal antibodies (Said et al. 2001,2004). Blots were striped and reprobed with anti β-actinantibodies (Santa Cruz, CA, USA) to normalize data forsample loading.

Determination of hTHTR-1 and hTHTR-2 promoteractivity: transient transfection and luciferase assay

We have previously described the full-length hTHTR-1and hTHTR-2 promoters fused to luciferase (Reidling& Said, 2003; Nabokina & Said, 2004). ARPE-19 cellswere cotransfected with Lipofectamine 2000 at ∼75%confluency with 2 μg of either full-length hTHTR-1 orhTHTR-2 reporter gene construct along with 100 ngof the pRL-TK (Renilla luciferase-thymidine kinase)(Promega) to normalize for transfection efficiency.Cell lysates were then prepared from cells 48 h aftertransfection, and luciferase activity was determined aspreviously described (Reidling & Said, 2003; Subramanianet al. 2003a; Nabokina & Said, 2004). Data are presented asfold increase in expression over pGL3 basic expression setarbitrarily at 1 as previously described (Reidling & Said,2003; Subramanian et al. 2003a; Nabokina & Said, 2004).

Generation of hTHTR-1–GFP mutants and confocalimaging of live ARPE-19 cells

The QuikChange site-directed mutagenesis kit(Stratagene, La Jolla, CA, USA) was used to introduce

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

76 V. S. Subramanian and others J Physiol 582.1

insertions or deletions of nucleotides into the openreading frame of hTHTR-1 (the product of SLC19A2gene). The hTHTR-1–GFP fusion protein was constructedby performing PCR using gene-specific primers (Table 1).PCR products and the GFP–N3 vector were digested withXhoI and BamHI restriction enzymes and the productswere isolated from the gel and ligated to generate anin-frame fusion protein (hTHTR-1–GFP), with GFPfused to the C-terminus of hTHTR-1, as previouslydescribed (Subramanian et al. 2003b). Paired sense- andanti-sense primer oligonucleotides encompassing thespecified mutation sites (Table 1), as well as a plasmidcontaining hTHTR-1 fused to GFP–N3 were used as atemplate for PCR mutagenesis. Nucleotide changes in allconstructs were confirmed by sequencing (Laragen, LosAngeles, CA, USA).

0 2 4 6 8 10

0

50

150

250

Time (min)

Thia

min

e u

pta

ke (

fmol/m

g p

rote

in) 15 nM

0

5

10

15

0 2 4 6 8 10

Time (min)

Thia

min

e u

pta

ke (

pm

ol/m

g p

rote

in)

5 μM

A

B

Figure 1. Time-dependent uptake of thiamine by confluentmonolayers of ARPE-19 cellsARPE-19 cells were incubated at 37◦C in Krebs–Ringer solution(pH 7.4) for 0–10 min in the presence of 15 nM (A) and 5 μM (B)thiamine. Values are means ± S.E.M. of n = 3–4 separate uptakedeterminations. Note, in all figures when not apparent, the error barsare smaller than the symbol.

For confocal imaging of clinically relevant hTHTR-1mutants in ARPE-19 cells, cell monolayers grown oncoverslip-fixed Petri dishes were imaged for constructexpression using a Nikon C-1 confocal scanner headattached to a Nikon inverted phase-contrast microscope.Fluorophores were excited using the 488 nm line from anargon ion laser, and emitted fluorescence was monitoredwith a 530 ± 20 nm band pass (GFP).

Data presentation and statistical analysis

All uptake determinations are the results of at leastthree separate determinations and are expressed asmeans ± s.e.m. in pmol or fmol per mg protein per unittime. Comparison was made relative to simultaneouslyperformed controls using ANOVA or Student’s t test(unpaired), with a significant P value set at < 0.05. Kineticparameters of the saturable component of thiamine uptakewere determined by subtracting the diffusing component(calculated from the slope of the line between uptake atby a pharmacological concentration of 500 μm and thepoint of origin, i.e. multiplication of the slope by individualconcentration) from total uptake; data were then applied toa computerized model of the Michaelis–Menten equationas previously described (Wilkinson, 1961). Confocalimaging, real-time PCR, semiquantitative RT-PCR andWestern blot analyses were performed and promoteractivities were determined on at least three separateoccasions.

Results

Overall characteristics of the thiamine uptake processby ARPE-19 cells: time course, effect of incubationtemperature, pH and Na+

Thiamine uptake by ARPE-19 cells was linear with timeduring 10 min incubation (rate, 220 fmol (mg protein)−1

and 13.5 pmol (mg protein)−1 for 15 nm and 5 μm,respectively; r = 0.99 for both; Fig. 1). The metabolic formof the radioactivity taken up by the confluent ARPE-19cells following incubation for 10 min with [3H]-thiamine(40 nm) was found (by cellulose-precoated TLC plates; seeMethods) to be (95%) in the form of intact thiamine.

The initial rate of uptake of thiamine (15 nm) variedwith pH (Fig. 2). The highest uptake was observed overa pH range of 7.5–8. Isosmotic replacement of Na+

by Li+ or mannitol did not affect the initial rate ofthiamine (15 nm) uptake (153 ± 3.0, 149 ± 0.5 and 148 ±1.0 fmol (mg protein)−1 (7 min)−1 in the presence of Na+,Li+ and mannitol, respectively). Pre-treating ARPE-19monolayers (for 30 min at 37◦C) with 1 mm ouabain (aNa+–K+-ATPase inhibitor) did not significantly affect theinitial rate of thiamine (15 nm) uptake (157 ± 6.1 and

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 77

145 ± 3.3 fmol (mg protein)−1 (7 min)−1 in the presenceand absence of ouabain, respectively). Finally, uptake ofthiamine (15 nm) was temperature dependent (153 ± 4.0,73 ± 5.0 and 28 ± 1.7 fmol (mg protein)−1 (7 min)−1 at37◦C, 22◦C and 4◦C, respectively).

Involvement of saturable processes in thiamineuptake by ARPE-19 monolayers

The initial rate of thiamine uptake as a function ofconcentration was examined over nanomolar (15–50 nm)and micromolar (0.1–10 μm) concentration ranges.Evidence for saturable uptake was observed for bothconcentration ranges (Fig. 3A and B). Kinetic parametersof both saturable components were determined as statedin the Methods. The apparent K m and V max of thesaturable process in the nanomolar range were 30 ± 5 nm

and 260 ± 21 fmol (mg protein)−1 (7 min)−1, respectively,and in the micromolar range were 1.72 ± 0.3 μm and9.3 ± 0.8 pmol (mg protein)−1 (7 min)−1, respectively.This suggested that the thiamine uptake process ofARPE-19 cells is carrier-mediated. To confirm this, wealso examined the effect of unlabelled thiamine and thatof its structural analogues amprolium, benfotiamineand oxythiamine (all at 10 μm) on the initial rate of[3H]-thiamine (2 μm) uptake. Unlabelled thiamine causedsignificant (P < 0.01) inhibition of [3H]-thiamine uptake(18.3 ± 1 and 6.4 ± 0.03 pmol (mg protein)−1 (7 min)−1

in controls and in the presence of unlabelled thiamine,respectively); similarly, the thiamine analoguescaused significant (P < 0.01 for all) inhibition of[3H]-thiamine uptake (19.2 ± 1, 7.5 ± 0.2, 8.7 ± 0.1and 8.6 ± 0.3 pmol (mg protein)−1 (7 min)−1 in controlsand in the presence of amprolium, benfotiamine

0

50

100

150

200

5.0 6.0 7.0 8.0

pH

Thia

min

e u

pta

ke(f

mol/m

g p

rote

in/7

min

)

Figure 2. Effect of incubation buffer pH on thiamine uptake byconfluent monolayers of ARPE-19 cellsARPE-19 cells were incubated at 37◦C for 7 min in Krebs–Ringerbuffer at pH 5.0–8.0. [3H]-Thiamine (15 nM) was added at the onset ofincubation. Data are means ± S.E.M. of n = 3–4 separate uptakedeterminations.

and oxythiamine, respectively). No effect of biotin(1 mm) and the organic cations tetraethylammonium(TEA) and N-methylnicotinamide (NMN) (both at100 μm) were observed on the uptake of [3H]-thiamine(15 nm) (155 ± 7.0, 156 ± 2.0, 152 ± 2.3 and154 ± 2.0 fmol (mg protein)−1 (7 min)−1 in controls andin the presence of biotin, TEA and NMA, respectively).

We also tested possible trans-stimulation of[3H]-thiamine transport across the ARPE-19 cell

0

2

4

6

8

10

Thia

min

e u

pta

ke(p

mol/m

g p

rote

in/7

min

)

0 3 6 9 12

Thiamine [μM]

0

50

100

150

200

Thia

min

e U

pta

ke(f

mol/m

g p

rote

in/7

min

)

0 20 30 40 50

Thiamine [nM]

A

B

Figure 3. Uptake of thiamine by confluent monolayers ofARPE-19 cells as a function of nanomolar and micromolarconcentrationsARPE-19 cells were incubated at 37◦C for 7 min in Krebs–Ringersolution at pH 7.4 in the presence of nanomolar (15–50 nM; A) andmicromolar (1–10 μM; B) concentrations of thiamine. Uptake plotsshown are those of the saturable components calculated as describedin the text. Values are means ± S.E.M. of n = 3–4 separate uptakedeterminations.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

78 V. S. Subramanian and others J Physiol 582.1

membrane by unlabelled thiamine. Here we firstpreloaded the cells with 15 nm [3H]-thiamine (for 10 minat 37◦C) then incubated the cells (for 10 min) in thepresence and absence of 100 μm unlabelled thiamine inthe incubation buffer. The results showed the cellularcontent of 3H radioactivity to be significantly (P < 0.01)lower in ARPE-19 cells incubated in the presence ofunlabelled thiamine compared to those incubated in itsabsence (cell content of 3H radioactivity was 78.7 ± 1.6and 112 ± 2.0 fmol (mg protein)−1, respectively).

Effect of pharmacological agents and sulfhydrylgroup inhibitor on thiamine uptake by ARPE-19 cells

The diuretic amiloride, and the anti-trypanosomal drugmelarsoprol and its metabolite melarsenoxide have beenreported to inhibit thiamine transport in certain humancell types as well as in yeast and bacteria (Said et al. 1999,2001; Schweingruber, 2004; Szyniarowski et al. 2006). Forthis reason we examined the effect of these compoundson initial rate of thiamine uptake by ARPE-19 cells.Amiloride was found to cause a concentration-dependentinhibition of thiamine uptake (157 ± 7.1, 90.6 ± 3.4,55.8 ± 1.7 and 38.9 ± 1.2 fmol (mg protein)−1 (7 min)−1

for control and in the presence of 0.1, 0.5 and1 mm amiloride, respectively). On the other hand,neither melarsoprol nor melarsenoxide (each at150 μm) showed an effect (119 ± 3.8, 115 ± 6.7 and120 ± 4 fmol (mg protein)−1 (7 min)−1 for control andin the presence of melarsoprol and melarsenoxide,respectively) on thiamine uptake by ARPE-19 cells.

The effect of pretreating (for 30 min) theARPE-19 cells with the sulfhydryl group inhibitorp-chloromercuriphenyl sulphonate (p-CMPS, 0.05 mm)on the initial rate of thiamine (15 nm) uptake wasalso tested. The results showed significant (P < 0.01)inhibition of thiamine uptake by p-CMPS (143 ± 2.3and 75.3 ± 0.7 fmol (mg protein)−1 (7 min)−1 for controland p-CMPS pretreated cells, respectively). Furthertreating the p-CMPS-treated ARPE-19 cells with 10 mm

of the reducing agent dithiothreitol (DTT) led to asignificant (P < 0.05) reversal of the inhibitory effectof pCMPS on thiamine uptake (143 ± 2.3, 75.3 ± 0.7and 102.2 ± 1.7 fmol (mg protein)−1 (7 min)−1 forcontrol, p-CMPS and p-CMPS plus DTT-treated cells,respectively).

Expression of the hTHTR-1 and hTHTR-2 in ARPE-19cells and in native human retina

Expression of the hTHTR-1 and hTHTR-2 at the RNA levelwas examined by means of semiquantitative and real-timePCR techniques. The semiquantitative RT-PCR (describedin the Methods) showed clear expression of hTHTR-1

and hTHTR-2 in ARPE-19 cells (Fig. 4A). The twotransporters are also expressed in native human retinaltissue (Fig. 4A). We also more accurately quantified therelative level of mRNA expression of hTHTR-1 andhTHTR-2 in ARPE-19 cells (Fig. 4B) and native humanretina (Fig. 4C) using real-time PCR (see Methods) andfound the level of expression of the hTHTR-1 to besignificantly (P < 0.01) higher than the level of expressionof hTHTR-2 (Fig. 4B and C). Expression of the hTHTR-1and hTHTR-2 at the protein level in ARPE-19 cells wasalso determined by means of Western blot analysis (seeMethods) with the results showing expression of bothtransporters. The level of hTHTR-1 protein, however,was significantly (P < 0.05) higher than that of hTHTR-2(Fig. 4D). We also tested the relative activity of theSLC19A2 and SLC19A3 promoters in ARPE-19 cells (seeMethods). Activity of both promoters was found to besignificantly higher than that of pGL3 basic (∼280-and ∼4-fold, for SLC19A2 and SLC19A3 promoters,respectively). Relative activity of the SLC19A2 promoterwas found to be significantly (P < 0.01) higher than thatof the SLC19A3 promoter (Fig. 4E).

Regulation of the thiamine uptake process ofARPE-19 cells by extracellular thiamine level and byintracellular regulatory pathways

Effect of extracellular thiamine levels. The effect ofmaintaining the ARPE-19 cells in growth medium(for 5 days) containing specific levels of thiamine(1 nm and 1, 12 and 100 μm) on the initial rateof uptake of [3H]-thiamine (15 nm) was examined.The results showed [3H]-thiamine uptake to beinversely correlated with the level of thiamine inthe growth medium (181 ± 6, 125 ± 2, 113 ± 3 and94 ± 2 fmol (mg protein)−1 (7 min)−1 in the presence of1 nm and 1, 12 and 100 μm thiamine, respectively). Uptakeof the unrelated biotin, however, was similar in cells grownin the presence of 1 nm and 12 μm thiamine (161 ± 8 and170 ± 1 fmol (mg protein)−1 (7 min)−1, respectively). Todetermine whether extracellular thiamine concentrationsaffect the level of expression of the hTHTR-1 andhTHTR-2 at the mRNA and protein levels, we performedreal-time PCR and Western blotting, respectively, onsamples obtained from cells grown in the presence of1 nm and 12 μm thiamine. The results showed a significant(P < 0.05) increase in the steady-state mRNA (Fig. 5Aand B) and protein levels (Fig. 5C and D) of both thehTHTR-1 and hTHTR-2 in ARPE-19 cells grown in thepresence of 1 nm thiamine compared to those grownin the presence of 12 μm thiamine. We also tested theeffect of growing the ARPE-19 cells in the presence of1 nm and 12 μm thiamine on activity of the full-lengthSLC19A2 and SLC19A3 promoters. The results showed

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 79

significantly (P < 0.05) higher SLC19A2 (Fig. 5E) andSLC19A3 (Fig. 5F) promoter activities in cells grown inthe presence of 1 nm compared to those grown in thepresence of 12 μm thiamine.

Role of intracellular signalling pathways. We examinedthe possible involvement of a number of intracellularsignalling regulatory mechanisms (Ca2+–CaM, proteinkinase C (PKC), protein kinase A (PKA), proteintyrosine kinase (PTK) and nitric oxide (NO)) inregulating thiamine uptake by ARPE-19 cells. Therole of the Ca2+–CaM-mediated pathway in theregulation of thiamine uptake by ARPE-19 cellswas tested by examining the effect of pretreatingconfluent ARPE-19 cells (for 1 h) with three differentmodulators of the Ca2+–CaM- mediated pathways, namely

hTHTR-2

ARPE-19 Human retinan

300

200

100

0

0

5

15

25

0

5

15

25

β-actin

hTHTR-1

SLC19A2 SLC19A3

Luci

fera

se a

ctiv

ity o

ver

pG

L3 b

asi

c (X

fold

)

β-actin

hT

HT

R-1

hT

HT

R-2

hTHTR1 hTHTR2 hTHTR1 hTHTR2

Rela

tive m

RN

A e

xpre

ssio

n

Rela

tive m

RN

A e

xpre

ssio

n

ARPE-19 Human retina

~80kDa

~43kDa

A B

D E

C

Figure 4. Expression of hTHTR-1 and hTHTR-2 in ARPE-19 cells at the mRNA and protein levels and onactivity of SLC19A2 and SLC19A3 promotersSemiquantitative RT-PCR products from ARPE-19 cells and native human retina for hTHTR-1 and hTHTR-2 mRNAlevels were analysed on an agarose gel as described in the Methods. Data shown are representative of three separatesets of experiments (A). Real-time quantitative PCR of hTHTR-1 and hTHTR-2 mRNA levels in ARPE-19 cells (B) andnative human retina (C). Data are from at least three independent sets of experiments and normalized againstβ-actin and then quantified using a relative relationship method supplied by the iCycles manufacturer (Bio-Rad, CA,USA) and as described in the Methods. Western blot analysis was performed using 150 μg membranous fractionsof ARPE-19 cells and specific polyclonal antibodies against hTHTR-1 and hTHTR-2 (D). SLC19A2 and SLC19A3promoter-luciferase activity was determined in ARPE-19 cells as described in the Methods. Values presented aremeans ± S.E.M. of three separate determinations performed on different occasions and are expressed as luciferaseacivity of each construct over the promoter-less vector pGL3 basic following transient transfection into ARPE-19cells. To account for transfection efficiency, Firefly luciferase activity was normalized relative to the activity ofsimultaneously transfected Renilla luciferase (D).

calmidazolium, trifluoperazine (TFP) and 1-[N, O-bis(5-isoquinolinesulphonyl)-N-methyl-l-tyrosyl]-4-phenylpiperazine (KN-62) on the initial rate of [3H]-thiamine(15 nm) uptake. All inhibitors tested were foundto cause significant (P < 0.01) inhibition ofthiamine uptake (124 ± 3, 64.6 ± 2, 75 ± 3.8 and48 ± 2 fmol (mg protein)−1 (7 min)−1 for control andin the presence 10 μm calmidazolium, 25 μm TFP and10 μm KN-62, respectively).

No role for the PKC-mediated pathway in the regulationof thiamine uptake by ARPE-19 cells was evident as uptakeof thiamine (15 nm) was not affected by pretreatmentwith modulators of this pathway (116 ± 3.8, 106 ± 0.34,112 ± 3.9 and 114 ± 10.2 fmol (mg protein)−1 (7 min)−1

for control and in cells pretreated with (all 10 μm)phorbol 12-myristate 13-acetate, staurosporin and

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

80 V. S. Subramanian and others J Physiol 582.1

chelerythrine, respectively). Similarly no role for thePKA-mediated pathway was evident as modulators ofthis pathway also failed to significantly affect thiamineuptake by ARPE-19 cells (123 ± 2.6, 118 ± 3.7 and120 ± 3.2 fmol (mg protein)−1 (7 min)−1 for control and

700

500

300

100

0

0

5

0

5

hTHTR-1

β-actinhTHTR-2

β-actin

Contr

ol

Defic

ient

DeficientControl DeficientControl

0

5

10

15

Luci

fera

se a

ctiv

ity o

ver

pG

L3 b

asi

c (X

fold

)

Luci

fera

se a

ctiv

ity o

ver

pG

L3 b

asi

c (X

fold

)

Rela

tive m

RN

A e

xpre

ssio

n

Control Deficient

Re

lativ

e m

RN

A e

xpre

ssio

n

Control Deficient

Contr

ol

Defic

ient

hTHTR-1 hTHTR-2

SLC19A2SLC19A3

~80kDa

~43kDa

~80kDa

~43kDa

A B

D

F

C

E

Figure 5. Effect of thiamine deficiency on hTHTR-1 and hTHTR-2 mRNA and protein levels and on activityof SLC19A2 and SLC19A3 promoters in ARPE-19 cellsReal-time PCR analysis of hTHTR-1 (A) and hTHTR-2 (B) mRNA levels in ARPE-19 cells grown under control (12 μM)and thiamine deficient (1 nM) conditions. Data are from at least three independent sets of experiments andnormalized against β-actin and then quantified as described in the Methods. Western blot analysis of hTHTR-1(C) (relative density of the bands were determined and are as follows: for hTHTR-1 the intensity was 946385and 1100047 and for β-actin the intensity was 707596 and 32997 for control and thiamine-deficient cells,respectively) and hTHTR-2 (D) proteins in membranous fractions of ARPE-19 cells grown under control (12 μM)and thiamine-deficient (1 nM) conditions. Data shown are representative of three separate sets of experiments.Activity of full-length SLC19A2 (E) and SLC19A3 (F) promoters in ARPE-19 cells grown in control (12 μM) andthiamine-deficient (1 nM) conditions. Data are means ± S.E.M. of three separate determinations performed ondifferent occasions.

following pretreatment with 0.5 mm dibutyryl cAMPor 8-bromo-cAMP, respectively). Similarly, no role forthe PTK-mediated pathway (as indicated by lack ofeffect of modulators of this pathway –genistein andtyrophospin A-25) or for the NO-mediated pathway (as

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 81

indicated by lack of effect of modulators of this pathway– S-nitrose-N-acetylpenicillamine, sodium nitroprussideand 8-bromo cGMP) in the regulation of thiamine uptakeby ARPE-19 cells was observed (data not shown).

Imaging of hTHTR-1 mutants found in TRMA in ARPE-19cells. TRMA is caused by mutations in hTHTR-1and is associated with visual disturbances and retinalabnormality (Fleming et al. 1999; Diaz et al. 1999; Labayet al. 1999; Raz et al. 2000; Meire et al. 2000; Scharfeet al. 2000). In the present study we resolved the cellularlocalization of three clinically identified TRMA mutantsin hTHTR-1 (D93H, S143F and G172D) by generatingthese mutants by means of site-directed mutagenesisfollowed by their transfection individually into ARPE-19cells and by performing live cell imaging of their cellulardistribution by means of confocal microscopy. Theresults showed wild-type hTHTR-1–GFP protein to betargeted predominantly to the plasma membrane; it wasalso evident within a variety of intracellular vesicularstructures (Fig. 6A). The individual mutants, however,showed different patterns of cellular expression (Fig. 6A).The D93H and S143F mutants were also expressed atthe cell surface with a lower level of expression efficiencythan that of the wild-type (hTHTR-1). By contrast, theG172D mutant was found to be completely retained withinthe endoplasmic reticulum (Fig. 6A). Further, all mutantsshowed significantly (P < 0.01) impaired functionalitywhen compared to wild-type hTHTR-1 followingtransfection into ARPE-19 cells (Fig. 6B).

Discussion

In this study we investigated the thiamine uptake processof hRPE using the ARPE-19 cell line as an in vitro modelsystem. This cell line possesses many of the structural andphysiological characteristics of hRPE in vivo and has beenextensively used in physiological investigations (Dunnet al. 1996; Aukunuru et al. 2001; Busik et al. 2002; Philpet al. 2003). Our aims in this study were: (i) to investigatethe physiological characteristics/mechanism of thiamineuptake; (ii) to assess the expression of the hTHTR-1 andhTHTR-2 in ARPE-19 and in native human retina; (iii) toexamine possible regulation of the thiamine uptake processby extracellular and intracellular factors; and (iv) to studythe effect of specific clinically relevant hTHTR-1 mutationson the expression of the hTHTR-1 transporter in live cells.All these issues are discussed in turn below.

The physiological characteristics/mechanismof the thiamine uptake process

Thiamine uptake by the ARPE-19 cells occurs withoutmetabolic alterations in the transported substrate, and the

initial rate of thiamine uptake is both temperature- andenergy-dependent in nature. Thiamine uptake was foundto be Na+-independent as replacing Na+ in the incubationbuffer or pretreating the cells with the Na+–K+-ATPaseinhibitor ouabain failed to affect thiamine uptake. Uptakeof thiamine, however, was pH-dependent and decreasedwith a decrease in the incubation buffer pH from 7.4to 5.0. These findings are similar to those reportedpreviously for thiamine uptake in epithelial cell typeswhere the thiamine uptake process was suggested to involvea thiamine–H+-exchange mechanism (Rindi & Laforenza,2000; Said et al. 2001).

The thiamine uptake process of the ARPE-19 cellsappears to involve two saturable systems, one beingfunctional in the nanomolar range with high affinity

0

100

300

500

700

Contro

lGFP

hTHTR1-

GFP

hTHTR-1

[D93

H]-GFP

hTHTR-1

[S14

3F]-G

FP

hTHTR-1

[G17

2D]-G

FP

Thia

min

e u

pta

ke(f

mol/m

g p

rote

in/7

min

) *

** ****

hTHTR-1-GFP hTHTR-1[D93H]-GFP

hTHTR-1[S143F]-GFPhTHTR-1[G172D]-GFP

A

B

Figure 6. Distribution of individual clinically relevanthTHTR-1–GFP mutations in ARPE-19 cellsA, representative of confocal lateral (xy) images showing localization ofindividual mutant constructs in ARPE-19 cells, 24–48 h after transienttransfection. B, uptake of [3H]-thiamine in control, GFP, hTHTR-1–GFP,hTHTR-1[D93H]–GFP, hTHTR-1[S143F]–GFP and hTHTR-1[G172D]–GFPtransiently expressing ARPE-19 cells. ∗Significantly different fromcontrol (P < 0.01); ∗∗no significant difference between mutants andcontrol. Results represents means ± S.E.M. from three separatedeterminants performed on three different occasions.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

82 V. S. Subramanian and others J Physiol 582.1

but low capacity, and the other being functional inthe micromolar range with lower affinity but highercapacity. This may be a reflection of the functionalityof the hTHTR-2 and hTHTR-1 (both of which areexpressed in ARPE-19 cells), respectively. This is becausethe apparent K m values of these thiamine transportsystems fall within the nanomolar and the micromolarconcentration ranges, respectively (Dutta et al. 1999;Said et al. 2004). The ability of unlabelled thiamine andits structural analogues amprolium, benfotiamine andoxythiamine to inhibit the uptake of [3H]-thiamine furtherconfirms the carrier-mediated nature of the thiamineuptake process of the ARPE-19 cells. Because uptake ofthe cationic thiamine was not affected by the presenceof the unrelated organic cations TEA and NMN in theincubation buffer, we conclude that the thiamine uptakeprocess in these cells is specific in nature. Finally theability of unlabelled thiamine in the incubation bufferto stimulate [3H]-thiamine efflux from preloaded cellsfurther confirmed the carrier-mediated nature of theuptake process.

We also examined the effect of the diuretic amiloride(which inhibits thiamine uptake in other human epithelialcells; Said et al. 1999, 2001; Ashokkumar et al. 2006)and the anti-trypanosomal drug melarsoprol and itsmetabolite melarsenoxide (which inhibit thiamine uptakeby neuroblastoma cells as well as by yeast and bacteria;Schweingruber, 2004; Szyniarowski et al. 2006) on theinitial rate of thiamine (15 nm) uptake. The resultsshowed that whereas melarsoprol and melarsenoxidedo not affect thiamine uptake, amiloride causes aconcentration-dependent inhibition of thiamine uptake.The former findings raise the possibility that the thiamineuptake process exhibits different pharmacologicalporperties in different cell types; the latter findings,on the other hand, reveal another way in whichamiloride appears to interfere with thiamine transportand points to the need for careful investigation of thepotential adverse effect of this pharmacological agent onbody/cellular thiamine homeostasis in vivo.

The uptake process of thiamine by ARPE-19 cells wasfound for the first time to be sensitive to the inhibitoryeffect of the SH group inhibitor p-CMPS. The ability of thereducing agent DTT to significantly reverse the inhibitoryeffect of p-CMPS on thiamine uptake by ARPE-19 cellssuggests possible involvement of such groups in the uptakeof thiamine.

Expression of hTHTR-1 and hTHTR-2 in ARPE-19 cellsand in native human retina

Previous studies have shown that the hTHTR-1 andhTHTR-2 are differentially expressed in variety of tissues,but their level of expression in the hRPE has not been tested

previously. This issue was addressed in this investigationby means of semiquantitative, real-time PCR and Westernblotting with the results showing expression of bothcarriers at both the mRNA and protein levels. Similarly,the two thiamine transporters were found for the firsttime to be expressed in native human retina. In both ofthe above cases, expression of the hTHTR-1 was foundto be significantly higher than that of the hTHTR-2, ascenario that has been seen previously in other humanepithelial cells (Said et al. 2004; Ashokkumar et al. 2006). Itis interesting that when activity of the full-length SLC19A2and SLC19A3 promoters were examined in ARPE-19 cells,activity of the former promoter was also found to besignificantly higher than that of the latter promoter, afinding that is in line with the level of mRNA expressionof the two carriers.

Regulation of thiamine uptake

Possible regulation of the thiamine uptake process ofthe ARPE-19 cells by extracellular and an intracellularfactors was also investigated in these studies. Performingsuch investigations is physiologically important as suchfactors affect nutrient/substrate transport in a cell-specificmanner (Traebert et al. 1999; Reidling & Said, 2005).Studies on the effect of extracellular thiamine level showedan inverse correlation between substrate uptake and itsprevailing level in the growth medium. MaintainingARPE-19 cells in a thiamine-deficient growth medium wasfound to lead to a specific and significant up-regulationof [3H]-thiamine uptake compared to cells grown in thepresence of high concentrations of thiamine. Thisinduction of thiamine uptake was found to be associatedwith an increase in the level of expression of the hTHTR-1and hTHTR-2 at the mRNA and protein levels as wellas in the level of activity of the SLC19A2 and SLC19A3promoters in cells maintained in the thiamine-deficientgrowth medium compared to those supplemented witha high dose of thiamine. The latter observation suggeststhe possible involvement of transcriptional regulatorymechanism(s) in this adaptive regulation of the thiamineuptake process in ARPE-19 cells. Further studies areneeded to delineate the cis- and trans-regulatory elementsthat are responsible for mediating this adaptive response.

Investigating the possible role of specific intracellularregulatory pathways in the regulation of the thiamineuptake process of ARPE-19 cells has revealed that the PKC-,PKA-, PTK- and NO-mediated pathways appear to haveno impact on thiamine uptake in ARPE-19 cells. However,a possible role for a Ca2+–CaM-mediated intracellularpathway appeared to exist. The latter conclusion is basedon the observations that modulators of this pathway causesignificant inhibition of thiamine uptake. It is interestingthat thiamine uptake by other epithelial cells also appears

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 83

to be regulated by a Ca2+–CaM-mediated pathway (Saidet al. 1999; Ashokkumar et al. 2006), thus suggesting thepossible existence of a common intracellular mechanismfor regulating thiamine uptake in different cell types.Further studies are needed to elucidate the molecularmechanism(s) through which the Ca2+–CaM-mediatedintracellular pathway exerts its effect on thiamine uptake.

Clinical TRMA mutants

At least 16 distinct mutations in the hTHTR-1 havebeen identified in TRMA patient, encompassing missense,nonsense and frame-shift mutations (Raz et al. 2000;Lagarde et al. 2004; Ricketts et al. 2006). In this study, wereport for the first time the cellular effects of three missensemutations (D93H, S143F and G172D) on membraneexpression and functionality of the protein in livingARPE-19 cells. The results showed different cellulartargeting phenotypes with mutants D93H and S143Fbeing expressed at the cell surface with a lower level ofexpression efficiency than wild-type, and G172D mutantexhibiting a completely impaired trafficking phenotypewith the protein being retained in the endoplasmicreticulum in a manner analogous to that obtained withclinical point mutations occurring within other nutrienttransporters (Martin et al. 1997; Stein et al. 2002).Furthermore, all mutants displayed impaired thiamineuptake.

From the discussion presented above, the questionmay arise as to how mutations in hTHTR-1 in TRMApatients lead to low thiamine supplies (and retinalabnormalities and visual disturbances) knowing that RPEcells can up-regulate hTHTR-2 when faced with thiaminedeficiency. The most likely explanation to this is that thehigh metabolic rates of the retina and other ocular tissuesrequire them to need large amounts of thiamine thatcould not be provided even by up-regulated hTHTR-2in patients exhibiting impaired expression/activity ofhTHTR-1 (which as shown here is the dominant thiaminetransporter in retina).

In summary, our results characterized, for the firsttime, the thiamine uptake process in ARPE-19 cellsand demonstrated the involvement of a specializedcarrier-mediated process. Both the hTHTR-1 andhTHTR-2 were found to be expressed in ARPE-19cells as well as in native human retina. The resultsalso show that the hRPE thiamine uptake process isadaptively regulated by the extracellular thiamine levelvia a mechanism that involves both the hTHTR-1 andhTHTR-2; the uptake process also appears to be regulatedby an intracellular Ca2+–CaM-mediated pathway. Further,clinically relevant mutations of the hTHTR-1 in TRMAwere found to impair cell surface expression or functionof the transporter in these cells, thus providing the

connecting link between transport physiology and diseasecondition.

References

Ashokkumar B, Vaziri ND & Said HM (2006). Thiamin uptakeby the human-derived renal epithelial (HEK 293) cells:cellular and molecular mechanisms. Am J Physiol RenalPhysiol 291, F796–F805.

Aukunuru JV, Sunkara G, Bandi N, Threson WB & KompellaUB (2001). Expression of multi-drug resistance-associatedprotein (MRP) in human retinal pigment epithelial cells andits interaction with BAPSG, a noval aldose reductaseinhibitor. Pharm Res 18, 565–572.

Berdanier CD (1998). Advanced Nutrition-Micronutrients. CRCPress, New York.

Busik JV, Olson LK, Grant MB & Henry DN (2002).Glucose-induced activation of glucose uptake in cells fromthe inner and outer blood–retinal barrier. Invest OphthalmolVis Sci 43, 2356–2363.

Calingasan NY, Gandy SE, Baker H, Sheu KF, Smith JD, LambBT, Gearhart JD, Buxbaum JD, Harper C, Selkoe DJ, PriceDL, Sisodia SS & Gibson GE (1996). Noval neuritic clusterswith accumulations of amyloid precursor protein andamyloid precursor-like protein 2 immunoreactivity in brainregions damaged by thiamine deficiency. Am J Pathol 149,1063–1071.

Calingasan NY, Gandy SE & Gibson GE (1997). Thiaminedeficiency alters APP but not presenilin-1 immunoreactivityin vulnerable brain regions. Neuroreport 8, 2631–2634.

Diaz GA, Banikazemi M, Oishi K, Desnick RJ & Gelb BD(1999). Mutations in a new gene encoding a thiaminetransporter cause thiamine-responsive megaloblasticanaemia syndrome. Nat Genet 22, 309–312.

Dunn KC, Aotaki-Keen AE, Putkey FR & Hjelmeland LM(1996). ARPE-19, a human retinal pigment epithelial cellline with differentiated properties. Exp Eye Res 62, 155–169.

Dutta B, Huang W, Molero M, Kekuda R, Leibach FH, DevoeLD, Ganapathy V & Prasad PD (1999). Cloning of thehuman thiamine transporter, a member of the folatetransporter family. J Biol Chem 274, 31925–31929.

Eudy JD, Spiegel O, Barber RC, Wlodarczyk BJ, Talbot J &Finnell RH (2000). Identification and characterization of thehuman and mouse SLC19A3 gene: a novel member of thereduced folate family of micronutrient transporter genes.Mol Genet Metab 71, 581–590.

Fleming JC, Tartaglini E, Steinkamp M, Schorderit DF, CohenN & Neufeld EJ (1999). The gene mutated in thiamineresponsive anaemia with diabetes and deafness (TRMA)encodes a functional thiamine transporter. Nat Genet 22,305–308.

Frederikse PH, Farnsworth P & Zigler JS Jr (1999). Thiaminedeficiency in vivo produces fiber cell degeneration in mouselenses. Biochem Biopys Res Commun 258, 703–707.

Hammes HP, Du X, Edelstein D, Taguchi T, Matsumura T, JuQ, Lin J, Bierhaus A, Nawroth P, Hannak D, Neumaier M,Bergfeld R, Giardino I & Brownlee M (2003). Benfotiamineblocks three major pathways of hyperglycemic damage andprevents experimental diabetic retinopathy. Nat Med 9,294–299.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

84 V. S. Subramanian and others J Physiol 582.1

Labay V, Raz T, Baron D, Mandel H, Williams H, Barrett T,Szargel R, McDonald L, Shalata A, Nosaka K, Gregory S &Cohen N (1999). Mutations in SLC19A2 causethiamine-responsive megaloblastic anaemia associated withdiabetes mellitus and deafness. Nat Genet 22,300–304.

Lagarde WH, Underwood LE, Moats-Staats BM & Calikoglu AS(2004). Novel mutation in the SLC19A2 gene in anAfrican-American female with thiamine-responsivemegaloblastic anemia syndrome. Am J Med Genet A 123,299–305.

Leevy CM & Baker H (1968). Vitamins and alcoholism.Am J Clin Nutr 21, 325–328.

Livak KJ & Schmittgen TD (2001). Analysis of relative geneexpression data using real-time quantitative PCR and the2(-delta delta C (T)) method. Methods 25, 402–408.

Martin MG, Lostao MP, Turk E, Lam J, Wright EM & KremanM (1997). Compound missense mutations in thesodium/d-glucose cotransporter result in trafficking defects.Gastroenterology 112, 1206–1212.

Meire FM, Van Genderen MM, Lemmens K & Ens-DokkumMH (2000). Thiamine-responsive megaloblastic anemiasyndrome (TRMA) with cone-rod dystrophy. OphthalmicGenet 21, 243–250.

Nabokina SM & Said HM (2004). Characterization of the5′-regulatory region of the human thiamin transporterSLC19A3: in vitro and in vivo studies. J Cell Physiol 287,G822–G829.

Philp NJ, Wang D, Yoon H & Hjelmeland LM (2003). Polarizedexpression of monocarboxlate transporters in human retinalpigment epithelium and ARPE-19 cells. Invest OphthalmolVis Sci 44, 1716–1721.

Pow DV (2001). Amino acids and their transporters in theretina. Neurochem Int 38, 463–484.

Rajgopal A, Edmondson A, Goldman D & Zhao R (2001).SLC19A3 encodes a second thiamine transporter ThTr2.Biochim Biopys Acta 1537, 175–178.

Raz T, Labay V, Baron D, Szargel R, Anbinder Y, Barrett T, RablW, Viana MB, Mandel H, Baruchel A, Cayuela JM & CohenN (2000). The spectrum of mutations, including four novelones, in the thiamine-responsive megaloblastic anaemia geneSLC19A2 of eight families. Hum Mutat 13, 37–42.

Reidling JC, Nabokina SM, Balamurugan K & Said HM (2006).Developmental maturation of intestinal and renal thiaminuptake: studies in wild-type and transgenic mice carryinghuman THTR-1 and -2 promoters. J Cell Physiol 206,371–377.

Reidling JC & Said HM (2003). In vitro and in vivocharacterization of the minimal promoter region of thehuman thiamin transporter SLC19A2. Am J Physiol CellPhysiol 285, C633–C641.

Reidling JC & Said HM (2005). Adaptive regulation ofintestinal thiamin uptake: molecular mechanism usingwild-type and transgenic mice carrying hTHTR-1 and -2promoters. Am J Physiol Gastrointest Liver Physiol 288,G1127–G1134.

Reidling JC, Subramanian VS, Dudeja PK & Said HM (2002).Expression and promoter analysis of SLC19A2 in the humanintestine. Biochim Biophys Acta 1561, 180–187.

Ricketts CJ, Minton JA, Ariyawansa I, Wales JK, Lo IF & BarrettTG (2006). Thiamine-responsive megaloblastic anaemiasyndrome: long-term follow-up and mutation analysis ofseven families. Acta Paediatr 95, 99–104.

Rindi G & Laforenza U (2000). Thiamine intestinal transportand related issues: recent aspects. Proc Soc Exp Biol Med 224,246–255.

Said HM, Balamurugan K, Subramanian VS & Marchant JS(2004). Expression and functional contribution of thehuman thiamin transporter-2 (hTHTR-2) in thiaminabsorption in the human intestine. Am J Physiol GastrointestLiver Physiol 286, G491–G498.

Said HM, Ortiz A, Kumar C, Chatterjee N, Dudeja P & Rubin S(1999). Transport of thiamine in human intestine:mechanism and regulation in intestinal epithelial cell modelCaco-2. Am J Physiol Cell Physiol 277, C645–C651.

Said HM, Ortiz A, Subramanian VS, Neufeld EJ, Moyer MP &Dudeja P (2001). Mechanism of thiamine uptake by humancolonocytes: studies with cultured colonic epithelial cell lineNCM 460. Am J Physiol Cell Physiol 277, C645–C651.

Said HM, Wang S & Ma TY (2005). Mechanism of riboflavinuptake by cultured human retinal pigment epithelialARPE-19 cells: possible regulation by an intracellularCa2+–calmodulin-mediated pathway. J Physiol 566, 369–377.

Saito N, Kimura M, Kuchiba A & Itokawa Y (1987). Bloodthiamine levels in outpatients with diabetes mellitus. J NutrSci Vitaminol (Tokyo) 33, 421–430.

Scharfe C, Hauschild M, Klopstock T, Janssen AJM,Heidemann PH, Meitinger T & Jaksch M (2000). A novalmutation in the thiamine responsive megaloblastic anaemiagene SLC19A2 in a patient with deficiency of respiratorychain complex I. J Med Genet 37, 669–673.

Schweingruber ME (2004). The melaminophenyl arsenicalsmelarsoprol and melarsenoxide interfere with thiaminmetabolism in the fission yeast Schizosaccharomyces pombe.Antimicrob Agents Chemother 48, 3268–3271.

Stagg AR, Fleming JC, Baker MA, Sakamoto M, Cohen N &Neufeld EJ (1999). Defective high-affinity thiaminetransporter leads to cell death in thiamine-responsivemegaloblastic anemia syndrome fibroblasts. J Clin Invest103, 723–729.

Stein M-P, Wandinger-Ness A & Roitbak T (2002). Alteredtrafficking and epithelial cell polarity in disease. Trends CellBiol 12, 374–381.

Subramanian VS, Chatterjee N & Said HM (2003a). Folateuptake in the human intestine: promoter activity and effectof folate deficiency. J Cell Physiol 196, 403–408.

Subramanian VS, Marchant JS, Parker I & Said HM (2003b).Cell biology of the human thiamine transporter-1(hTHTR1): intracellular trafficking and membrane targetingmechanisms. J Biol Chem 278, 3976–3984.

Szyniarowski P, Bettendorff L & Schweingruber ME (2006).The antitrypanosomal drug melarsoprol competitivelyinhibits thiamin uptake in mouse neuroblastoma cells. CellBiol Toxicol 22, 183–187.

Tallaksen CME, Bohmer T & Bell H (1992). Blood and serumthiamin and thiamin phosphate esters concentrations inpatients with alcohol dependence syndrome before and afterthiamin treatment. Alcohol Clin Exp Res 16, 320–325.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society

J Physiol 582.1 Thiamine uptake by retinal pigment epithelia 85

Tanphaichirt V (1994). Thiamin. In Modern Nutrition in Healthand Disease, ed. Shils ME, Olsen JA & Shike M, pp 359–375.Lea & Febiger, New York.

Tomasulo PA, Kater RMH & Iber FL (1968). Impairment ofthiamine absorption in alcoholism in alcoholism. Am J ClinNutr 21, 1341–1344.

Traebert M, Hattenhauer O, Murer H, Kaissling B & Biber J(1999). Expression of type II Na-P(i) cotransporter inalveolar type II cells. Am J Physiol Lung Cell Mol Physiol 277,L868–L873.

Victor M, Adams RD & Collins GH (1989). TheWernicke–Korsakoff Syndrome and Related NeurologicalDisorders due to Alcoholism and Malnutrition. Davis,Philadelphia, PA.

Wilkinson GN (1961). Statistical estimation in enzyme kinetics.Biochem J 80, 324–332.

Acknowledgements

We would like to thank Mrs Shuling Wang for her excellent

technical assistance in this work. This study was supported by

grants from the National Institutes of Health (DK58057 and

DK56061 (H.M.S.) and DK 71538 (V.S.S.)) and the Department

of Veterans Affairs.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society